RESEARCH

Publications

The Spline GARCH Model for Low Frequency Volatility and its Global Macroeconomic Causes, (coauthored with Robert F. Engle), Review of Financial Studies, 21, 1187-1222.

Working Papers

The Factor-Spline-GARCH Model for High and Low Frequency Correlations (coauthored with Robert F. Engle)

We propose a new approach to model and estimate the correlation structure in equity markets. We characterize high and low frequency components of the correlation matrix by combining a factor model with other specifications capturing dynamic properties of volatilities and covariances between a single common factor and idiosyncratic returns. Spline-GARCH and DCC models describe the dynamic behavior of these terms. We examine the relevance of the underlying assumptions in a simple one-factor CAPM model through an empirical evaluation of a number of correlation specifications from a range of factor models. Adding dynamic properties to the idiosyncratic terms suggests remarkable improvements in fitting the data. Our factor specification is consistent with the behavior of correlations in the US market. It also shows a strong performance in anticipating correlations, especially at long horizons.

Macroeconomic Announcements, Price Discovery, and Order Flow Effects in the Stock Market: Evidence from Daily Data and Multiple Financial Markets.

This paper investigates heterogeneity in the market assessment of public macroeconomic announcements by exploring (jointly) two main mechanisms through which macroeconomic news might enter stock prices: Instantaneous fundamental news impacts consistent with the asset pricing view of symmetric information, and permanent order flow effects consistent with a microstructure view of asymmetric information related to heterogeneous interpretation of public news. Theoretical motivation and empirical evidence for the operation of both mechanisms are presented. Significant instantaneous news impacts are detected for news related to real activity (including employment), investment, inflation, and monetary policy; however, significant order flow effects are also observed on employment and FOMC announcement days. A multi-market analysis suggests that these asymmetric information effects come from uncertainty about future behavior of interest rates and inflation.

News, Announcements and Stock Market Volatility Dynamics.

This paper examines announcement and news effects on stock market volatility in the context of public disclosure of information regarding fundamental macroeconomic variables, monetary policy decisions, and earnings. The return distribution is parametrized using a jump Poisson-Gaussian model with time varying arrival intensity, and a GARCH model. Information surprises and announcements affect conditional volatility through a non-linear channel described by the jump intensity. We find that the day of the announcement, per se, has little impact on jump intensities. However, one exception is the day of employment releases where the jump intensity is significantly large. In contrast, when the surprise component of the announcement is incorporated in the model, inflation shocks show persistent effects and monetary policy shocks show short-lived effects. The jump model provides evidence of heterogeneous volatility persistence and asymmetric effects between positive and negative shocks from different types of news releases.

Research in Progress

Correlation Structure in Global Equity Markets, (coauthored with Robert F. Engle)

High and Low frequency correlations are estimated for 40 countries including developed economies, emerging markets and transition economies using daily data over the last 25 years. The empirical analysis sheds more light on how the correlation structure within developed and developing markets (as well as between them) has changed with financial globalization. Moreover, it examines the scope of this evolution in periods of financial distress compared to periods of low volatility.